HNHacker News
TopNewBestAskShowJobs

artursapek

8,170 karma · joined December 20, 2010

art@art.cx

https://art.cx

https://revise.io

submissionscomments
artursapek··on .self: A new top-level domain designed to support self-hosting
The .ai TLD is some tiny island with a few thousand people
artursapek··on "No, I swear I wrote this."
Trivial to simulate basic keystrokes. But I don't think it's trivial to simulate the natural process of drafting something. There's no concrete heuristic or algorithm (yet) for judging these types of replays, but I'd be impressed if someone can actually make a program that reliably simulates a natural keystroke-by-keystroke thought process which appears human when replayed in this way.
artursapek··on Previewing GPT‑5.6 Sol: a next-generation model
They claim extreme performance on ExploitBench, which Mythos was touted as being incredible at. https://x.com/OpenAI/status/2070555278576439306
artursapek··on Claude Fable 5
Fable 5 beats GPT 5.5 in my proofreading benchmark. And it does so at approximately the same total cost; it used significantly fewer turns than 5.5

https://x.com/tmuxvim/status/2064452096800198930

artursapek··on Siri AI
I would expect Apple to hedge their bet on Gemini and build everything so that the model can be swapped out in the future.
artursapek··on Apple reveals new AI architecture built around Google Gemini models
I use Carplay all the time and I didn't even realize it has voice control. I just set things up on my phone and drive.
artursapek··on Siri AI
I think it's fair to say that OpenAI has at least partially won the "consumer AI" segment.
artursapek··on The desperation of NYTimes
You’re not responding to anything the parent said.
artursapek··on Backpressure is all you need
I agree. I have gotten an incredible amount of work done iterating with 5-30 minute long agent tasks. But it requires I stay engaged, and not go chill on the beach, which I guess is a lot of agentmaxxers’ goal.
artursapek··on Backpressure is all you need
Yeah I don’t really see the backpressure analogy here - it implies that the agent is constantly producing new stuff, which isn’t really possible since the solution is very detailed specs/goals.
artursapek··on Can we have the day off?
Salaried people think they get paid for their output. No silly, you get paid for your time. Just like an hourly worker.
artursapek··on I'm Tired of Talking to AI
am I the only one who always replies to people manually? I don't think I've ever done the "send back a chatGPT screenshot" or copy paste a response from chatGPT to a message I know was from a human.
artursapek··on Not alive, but not dead: disembodied human brains used for drug testing
A nice reminder to not check the "organ donor" box at the DMV
artursapek··on Gaussian Splat of a Strawberry
ew the bottom is moldy
artursapek··on Local AI needs to be the norm
I'm someone who is trying to build a subscription-based business to cover underlying LLM costs, and very hopeful I can one day just sell a permanent license to the software instead with customers using local LLMs to power it.
artursapek··on Valve releases Steam Controller CAD files under Creative Commons license
Valve practically has a monopoly on PC gaming, I think they're pretty fat and happy too ;D
artursapek··on Belgium stops decommissioning nuclear power plants
I never said it was exclusively labor and energy... but those are prerequisites for producing anything like clothing, food...
artursapek··on Spell-Checking with LLMs
I'm eagerly awaiting that day! I want to ship a fully local version of Revise.
artursapek··on Belgium stops decommissioning nuclear power plants
Damn, that’s heavy man. I’m sorry. I don’t know your situation but men are fortunate enough to be able to reproduce later in life so you could still turn it around.

I had my first kid accidentally in college and dropped out to focus on that. Very grateful for it.

artursapek··on Belgium stops decommissioning nuclear power plants
Those are not mission critical for the survival of a civilization.
artursapek··on Belgium stops decommissioning nuclear power plants
The OP was drawing a connection between how the West has become less energy independent (not producing nuclear, importing energy) and how it’s become less labor independent (not producing people, importing them instead). The two are related because they are both caused by complacency, and they’re both destabilizing to the West.
artursapek··on Belgium stops decommissioning nuclear power plants
Having kids and raising them is your primary purpose as a man. Anything else you spend your time on is secondary to that.
artursapek··on Belgium stops decommissioning nuclear power plants
It’s absolutely related. It's constantly being excused by politicians as a solution to labor shortages.
artursapek··on Bankruptcies increase 11.9 percent
It's so disgusting watching how much sports gambling and "prediction markets" have exploded in recent years
artursapek··on An AI agent deleted our production database. The agent's confession is below
if your prod DB can be nuked with a single curl command, you are the problem
artursapek··on Show HN: I benchmarked how good LLMs are at proofreading English
Good idea about the leaderboard for open vs closed models!

Point taken on using an agent. I went that route because part of the goal for this benchmark is to inform which models I push in my agentic word processor, which uses tools for focused proofreading/editing. It's much faster and generally cheaper to use tools for surgical changes on large documents, rather than having the model spit out the entire document with all issues corrected. So yes, I am trying to measure agentic abilities here.

A simple one-pass full-rewrite test would also make an interesting benchmark, though.

artursapek··on GPT 5.5 sets new record in proofreading benchmark
Hi HN - this is a benchmark I developed that tests various models against large samples of text, asking them to find and fix a variety of errors. Its purpose is to evaluate how good models are at proofreading (a common use case of LLMs) and how efficient they are on various axes.

I've been working on this to inform my own decisions about which models to use in my agentic word processor, but I think it's also just useful data.

I just ran GPT 5.5 and it broke Gemini's previous high score of 92.5%!

The code and run artifacts are available on Github: https://github.com/reviseio/errata-bench

artursapek··on France confirms data breach at government agency that manages citizens' IDs
this is exactly my problem with them
artursapek··on Aliens.gov will be running as a WordPress multisite
how tf do you know they’re useless
artursapek··on Laravel raised money and now injects ads directly into your agent
Avoid VC funding at all costs
← PreviousPage 2 of 34Next →